This subproject is one of many research subprojects utilizing the resources provided by a Center grant funded by NIH/NCRR. The subproject and investigator (PI) may have received primary funding from another NIH source, and thus could be represented in other CRISP entries. The institution listed is for the Center, which is not necessarily the institution for the investigator. Biomedical Problem Cardiac arrhythmias and fibrillation are potentially life-threatening conditions that can result from improper conduction of electrical impulses in the heart. The disruption of electrical impulses impairs the mechanical pumping of blood and hence can quickly result in brain damage and death. It is therefore important to: 1) detect when and where problems are occurring, 2) understand the conditions under which arrhythmias and fibrillation arise, and 3) develop therapies to cure or reducing the episodes of arrhythmia. The scientific questions of detection, basic mechanisms and therapies have historically been asked and answered by experimentalists. Experimental studies, however, are costly, are limited by the quantities that can be measured, often have low spatial and temporal resolution, and in humans are very dangerous. In order to ease investigation, researchers are turning to computer simulations in order to study these phenomena in more depth. Computational Problem The goal of computational cardiac electrophysiology is to obtain preliminary data to guide fewer, more targeted experimental studies. The computational requirements for such simulations are severe. Many millions of computational elements must be evaluated to "step" the solution ahead in time. Those computations must be repeated several thousand times in order to calculate 1 millisecond of simulation time. For realistic simulations, several seconds or minutes must be simulated. A simple estimate indicates that for a whole human heart simulation roughly 1 heart beat would require hundreds of GB of memory and thousands or millions of CPU hours. This has historically necessitated simplifications in the size and complexity of the problem. Computer simulations of cardiac electrophysiology have recently become more realistic. The inclusion of realistic membrane dynamics, anatomically accurate geometries, and spatial inhomogeneities all increase the computational complexity of the simulation, as well as increasing the hardware resources required by the simulator. Furthermore, as the simulation complexity increases, software maintenance problems can be compounded;it is more difficult to detect what feature in the simulator caused a given bug. These challenges have led to the development of our modular simulation system, CardioWave